Tech Company’s AI Reporter Helps Meeting Attendees Be in Two Places at Once

Siemens AG used AI to spread key insights from an internal leadership meeting to participants — and remote global employees — in record time.

Author: Barbara Palmer       

Every year, Siemens AG, the global technology company, gathers its top leaders from around the world for an internal meeting it calls the Siemens Business Conference. At last October’s event in Munich, Germany, where the agenda included main stage presentations plus five breakout stages, Stephen Rose, Siemens AG’s head of global communication services, and his colleagues faced a challenge most meeting organizers would love to have.

Stephen Rose

Stephen Rose

All of the sessions offered such valuable content that the meeting’s 500 participants found it hard to decide which to attend, Rose told Convene. “Just imagine that you are at the conference and only able to participate in three out of five sessions, but all of them are important to you,” he said. And not only that, he added, but “hundreds of people elsewhere in the world are also waiting to hear from you about what was said at the sessions.” Rose’s solution: An artificial intelligence–assisted process that delivered summaries and key takeaways from all the sessions to meeting participants, beginning while the sessions were still in progress.

Rose and a team of 10-15 worked on the project for six weeks before the event, he said, using customized AI software on a fast-tracked timeline that included managing data security and legal issues, as well as technology. They used the event’s meeting app to channel content to participants through a character named “Sparky the AI reporter,” which had its own tile in the event app. Participants could download the summaries and takeaways on their phones and forward them to remote colleagues.

Pair AI With Human Intelligence

Before content from the sessions went up on the event app, it first went through Sina Auer, an editor on Rose’s team who was familiar with the material and could screen the output for accuracy. The editor was on the lookout for what Rose called “reputational risks.” For example, if an executive mentioned a circumstance that could potentially negatively influence revenue, AI could interpret that as a negative revenue prediction, Rose said. One of the key takeaways from the project was that artificial intelligence must be paired with natural — or human — intelligence. “You wouldn’t post anything on LinkedIn that someone else wrote for you without reading it,” he said. “If you have output from an AI tool, treat it as a draft and not as a final document.”

The feedback from the event’s speakers about the quality and accuracy of the content summaries and takeaways was, in general, very good, Rose said. The response from most, he said, was an incredulous “How did you do this?” Others were more critical, which Rose attributed to a mismatch between the message the speaker intended to convey and how it came across. Since then, with more use, the AI tool has become better able to recreate the tone of Siemens communications, Rose said. And, of course, he said, “the ‘natural intelligence’ also improves, the more one uses the AI tool.”

The positive reaction from participants, as well as the usage of the “Sparky” feature on the app, exceeded his expectations, Rose said. Meeting participants greatly appreciated the ability to compare how the content summaries matched up with their own understanding of a session’s key points as they were walking out the door, he said. One of the unanticipated benefits of the project was that the AI tool also delivered accelerated feedback to the event organizers and company executives, Rose said. He and his team plugged all the capturable content from the event that they could into the data, including audience polling survey results, comments about the sessions on the company’s internal social-media platforms, and questions that the participants asked during the sessions. Within an hour after the event’s conclusion, Rose and his team were already evaluating feedback from the event and providing analysis to the company’s CEO. There was even data, Rose said, about improvements they could make for the next event.

‘Ideas and Ideas and Ideas’

Another key takeaway from the project, Rose said, was the necessity of managing expectations around AI itself. “There is a lot of fantasy around AI and what it can do. Once we started this endeavor, people began having ideas and ideas and ideas about what else it could do.” The focus should not be on the functionality of the tool but the value you are adding with the tool — which, in this case, he said, was timely preparation of content for efficient communication with participants.

The next challenge will be to expand the use of the tool, he said. “There is so much potential.”

An AI-powered robot named Ameca co-presented part of the Welcoming Session at the NAB (National Broadcasters Association) 2024 in Las Vegas, where artificial intelligence in many forms was a hot topic.

It’s an educational task as much as a technological one, he said, to help people understand what AI software is — and isn’t — good for. “We are really keen on finding new, efficient, and realistic ways to really accelerate AI-supported communications and especially to make it accessible.”

Rose also is looking at the expanded possibilities of the tool beyond an event’s audio stream. “Maybe the next step is looking at the picture and video tracks, so that in the future we can also create more efficiency in the post-production of videos” using AI, Rose said. “I could say, ‘Show me all the scenes where we have people cheering,’ or ‘Show me all the scenes where the CEO talks about sustainability.’”

Rose advised event professionals to think about AI in terms of how it can be embedded in their work processes, not as a standalone tool. And he reinforced that the use of AI is inevitable. “I think there’s no way for us to not use it, because it will shape job roles and job profiles in the future,” he said. “And there will be more coming.”

Barbara Palmer is deputy editor of Convene.

From the Convene Podcast

Siemens’ Stephen Rose talked with Magdalina Atanassova, Convene’s digital media editor, about the future of AI on the Convene Podcast, Season 1, Episode 6, “Technology With Purpose Is the Key Word.” The conversation, recorded in September 2023 at Convening EMEA in Copenhagen, Denmark, was also joined by Albert Cerezales Garcia, a strategic consultant at MCI Group.

Become a Member

Get premium access to provocative executive-level education, face-to-face networking and business intelligence.